httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Pietro Cagnoni <pcagn...@mclink.net>
Subject Re: Memory leak issue?
Date Mon, 28 Jan 2002 09:16:19 GMT
> Background: I'm running a fresh build of Apache 1.3.22 (no external
> modules) on Slackware 8.0, Linux kernel 2.4.5.  This server has a P-III
> processor running at 1.0 GHz and 1.0 GB of RAM.  This machine has
> T1-equivalent bandwidth (1.5 Mbps) available to it, and serves about 20
> virtual hosts.  This server is primarily dedicated to Apache (although it
> it also runs a very low-usage PostgreSQL daemon and tinydns/dnscache.)
> 
> My problem is this: every 12 to 36 hours, I am paged with a frantic "the
> site is down!" message.  More precisely, connections are accepted, but the
> clients (or users) time out waiting for any data to be served.  Each time
> this happens, I observe the following:
> 
> 1. The total number of Apache child processes is equal to my MaxClients
> setting, as verified by "ps -ef | grep httpd | wc -l".
> 
> 2. The machine's RAM is completely full, and a good deal of my swap space
> is in use as well, as verified by "free -mo".
> 
> 3. The machine's CPU is hardly used at all, as verified by "top -d 1".
> 
> 4. Issuing an "apachectl graceful" takes care of the problem; all the
> virtual hosts immediately return to full speed.
> 
> I don't think that I need to upgrade this machine's RAM -- 1.0 GB should
> be more than sufficient for the task at hand.  I also don't want to
> periodically HUP Apache -- there has to be a better way.  What am I
> missing here?
> 
> If it makes any difference: we just deployed an e-commerce application,
> developed in-house in Perl, on a very busy vhost.  I thought that Perl
> automatically reclaims data-structure memory.  Could this have anything to
> do with it?
> 
> I'd be happy to make my httpd.conf file available, but I really didn't
> change it that much from the default.

try enabling the server-status url, with ExtendedStatus On, and check it
every now and then. you should be able to see the status of the "dead"
httpd daemons.

hope it helps.

pietro.

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Mime
View raw message